Abstract
UDP-glycosyltransferases (UGTs) are an important group of enzymes that participate in phase II metabolism of xenobiotics and use the cofactor UDP-glucuronic acid for the production of glucuronides. When acting on molecules bearing a carboxylic acid they can form acyl glucuronides, a group of metabolites that has gained significant interest in recent years because of concerns about their potential role in drug toxicity. In contrast, reports about the production of drug acyl glucosides (which might also display high reactivity) have been scarce. In this study, we discovered the formation of acyl glycoside metabolites of R- and S-ibuprofen (Ibu) by human liver microsomes supplied with the cofactor UDP-glucose. Subsequently, human UGT2B7*1 and UGT2B7*2 recombinantly expressed in fission yeast Schizosaccharomyces pombe could be shown to catalyze these reactions. Moreover, we could enhance the glucoside production rate in fission yeast by overexpressing the fission yeast gene SPCC1322.04, a potential UDP-glucose pyrophosphorylase (UGPase), but not by overexpression of SPCC794.10, and therefore suggest to name this gene fyu1 for fission yeast UGPase1. It was interesting to note that pronounced differences between the two polymorphic UGT2B7 variants were observed with respect to acyl glucoside production. Finally, using the metabolic precursor [13C6]glucose, we demonstrated the production of stable isotope-labeled reference standards of Ibu acyl glucoside and Ibu acyl glucuronide by whole-cell biotransformation in fission yeast.
Introduction
The UDP-glycosyltransferases (UGTs) are a superfamily of enzymes that catalyze the addition of glycosyl residues to small molecular-weight lipophilic chemicals and in this way play an important role in the elimination of different endogenous and exogenous compounds from the human body (Mackenzie et al., 2005). Humans have 22 UGTs that belong to four families: UGT1, UGT2 (with the subfamilies 2A and 2B), UGT3, and UGT8. The 19 human UGT isoforms in the UGT1 and UGT2 families have an often overlapping but sometimes very distinct substrate selectivity and are thought to typically prefer UDP-glucuronic acid (UDP-GA) as a sugar donor. UGTs may convert carboxylic-acid-containing drugs into acyl glucuronides (AGs), which because of their special properties and risks are of considerable pharmacological interest. It was noted decades ago that AGs are potentially reactive electrophiles that can interact with and covalently bind to nucleophilic targets (Faed, 1984). Accordingly, it was suggested that glucuronidation is not always a harmless detoxication reaction but that in some instances it can be a bioactivation pathway leading to potential toxicity (Spahn-Langguth and Benet, 1992). However, the overall toxicological significance of AGs is still a subject of debate (Boelsterli, 2011; Stachulski, 2011).
The nonsteroidal anti-inflammatory drug ibuprofen (Ibu) acts as an cyclooxygenase inhibitor and is predominantly orally applied as a racemic mixture; in the body, much of the R-enantiomer is converted to the active S-form (Lee et al., 1985). Ibu metabolism comprises a complex interaction of different phase I and phase II enzymes including several cytochrome P450 enzymes, dehydrogenases, and UGTs (Adams, 1992; Spraul et al., 1993; Kepp et al., 1997; McGinnity et al., 2000; Hao et al., 2005; Chang et al., 2008). To the best of our knowledge, all UGT metabolites of Ibu described so far are glucuronides, with the glucuronic acid being either attached to the acyl moiety of the parent compound (yielding Ibu AGs) or to hydroxy groups newly created in phase I metabolism.
In this study, we describe new acyl glycoside metabolites of Ibu. In vitro studies with human liver microsomes (HLMs) and the substrates (R)-(−)-ibuprofen (R-Ibu), (S)-(+)-ibuprofen (S-Ibu), and racemic ibuprofen (rac-Ibu) demonstrated that such metabolites are only produced in the presence of the cofactor UDP-glucose (UDP-Glc), but not UDP-galactose (UDP-Gal), thus confirming their identity to be either the R- or S-enantiomers of Ibu acyl glucoside (Ibu-Glc). UGT2B7 is one of the most important human UGT isoforms with respect to drug clearance in general (Williams et al., 2004) and significantly contributes to Ibu metabolism (Sakaguchi et al., 2004; Kuehl et al., 2005). There are two important polymorphic variants of this enzyme, UGT2B7*1 (268His) and UGT2B7*2 (268Tyr), with an allelic distribution of approximately 1:1 in white persons and 3:1 in Japanese individuals (Bhasker et al., 2000). Several studies demonstrated notable differences in the catalytic activity of these two forms, although this effect seems to be limited to certain substrates (Bernard et al., 2006; Thibaudeau et al., 2006; Belanger et al., 2009). Like some other UGT isoforms, UGT2B7 has been shown to use other UDP-sugars in addition to UDP-GA, such as UDP-Glc and UDP-Gal, and at least in some cases this property also seems to be substrate dependent (Mackenzie et al., 2003; Tang et al., 2003; Tang and Ma, 2005). We therefore investigated the ability of recombinantly expressed UGT2B7*1 and UGT2B7*2 to produce Ibu-Glc using our previously described fission yeast system (Dragan et al., 2010). Whole-cell biotransformations with UGT2B7-expressing fission yeasts demonstrated that the rate of metabolization of Ibu to Ibu-Glc depends on the Ibu enantiomer and the enzyme variant. Moreover, overexpression of an endogenous UDP-glucose pyrophosphorylase (UGPase) significantly enhanced Ibu-Glc production by UGT2B7*1, but not by UGT2B7*2, thus indicating a differential effect of cofactor availability. We suggest to name this UGPase, which is described in this study for the first time, fyu1, for fission yeast UGPase 1.
Materials and Methods
Chemicals.
R-Ibu was purchased from Enzo Life Sciences, Inc. (Farmingdale, NY), S-Ibu was purchased from Thermo Fisher Scientific (Waltham, MA), and rac-Ibu disodium salt was purchased from Sigma-Aldrich (St. Louis, MO). UDP-GA trisodium salt was purchased from Sigma-Aldrich, UDP-Glc disodium salt was purchased from Merck (Darmstadt, Germany), and UDP-Gal disodium salt was purchased from VWR (West Chester, PA). [13C6]glucose was purchased from Euriso-Top (Saint-Aubin, France). Methanol (high-performance liquid chromatography grade) was purchased from VWR. All other chemicals used were either from Carl Roth (Karlsruhe, Germany) or Sigma-Aldrich.
In Vitro Metabolism of Ibu Using HLMs.
Pooled HLMs (BD Biosciences, San Jose, CA) were incubated using UDP-GA, UDP-Glc, or UDP-Gal as cofactor as indicated. In addition, incubations without a UDP-sugar cofactor were done as a negative control. The total reaction volume was 100 μl. Final concentrations of in vitro incubations were 0.5 mg · ml−1 protein, 50 mM Tris HCl pH 7.5, 10 mM MgCl2, and 200 μM ibuprofen (S-Ibu, R-Ibu, or rac-Ibu). Reactions were started by the addition of 2 mM UDP-GA, UDP-Glc, or UDP-Gal, and samples were incubated for 18 h at 37°C and then mixed with an equal volume of acetonitrile. All experiments were done in triplicate.
Coding DNA Sequences.
The cDNA for human UGT2B7*2 optimized for expression in S. pombe was synthesized by Entelechon GmbH (Regensburg, Germany). The cDNA of human UGT2B7*1 was generated by site-directed mutagenesis using UGT2B7*2 as a template. The cDNAs of SPCC794.10 and SPCC1322.04 (fyu1) were amplified by polymerase chain reaction using cells of the S. pombe strain NCYC2036 (Losson and Lacroute, 1983) as a template. For amplification and introduction of 5′-NdeI- and 3′-BamHI sites, the following oligonucleotides were used: SPCC794.10, 5′-TAG TTT CAT ATG TTG CAT CGT CGA ATT C-3′ and 5′-TTT GGA TCC TCA ACA CTC CAT TAT TTT AC-3′; and SPCC1322.04, 5′-TAG ATA CAT ATG GAT TTG GCA CC-3′ and 5′-TTG GAT CCT TAG TGC TCC AAG ATA TTG-3′.
Media and General Techniques.
General DNA manipulating methods as well as media and genetic methods for fission yeast have been described (Sambrook and Rusell, 2001; Forsburg and Rhind, 2006). In addition, Edinburgh minimal medium (EMM) containing 100 g · l−1 glucose was used for biotransformation assays and EMM with 20 g · l−1 [13C6]glucose was used for the synthesis of isotope-labeled glucuronides.
Construction of Fission Yeast Strains.
Fission yeast strain construction was in principle done as described previously (Dragan et al., 2010). UGT2B7*1 and UGT2B7*2 cDNAs were cloned via NdeI and BamHI into the integrative vector pCAD1 (Dragan et al., 2005), which disrupts the leu1 locus of S. pombe and contains an ura4 marker gene. The construction of the uridine diphospho-glucose-6-dehydrogenase (UGDH) expression plasmid pREP1-UGDH has been described previously (Dragan et al., 2005). UGT2B7*1, UGT2B7*2, SPCC794.10, and SPCC1322.04 (= fyu1) cDNAs were cloned into the autosomal expression vector pREP1 using NdeI and BamHI, respectively. The correctness of all constructs was verified by automatic sequencing (Eurofins MWG Operon, Huntsville, AL or Seq-it GmbH, Kaiserslautern, Germany). pCAD1 and pREP1 contain the strong endogenous nmt1 promoter that permits regulation of expression via the presence or absence of thiamine in the media. Therefore, transformed cells were grown on EMM dishes with 5 μM thiamine to allow for better growth under repressed conditions.
All strains used in this study are listed in Table 1. Plasmids pCAD1- UGT2B7*1 and pCAD1-UGT2B7*2 were prepared as described previously (Dragan et al., 2005) and used for the transformation of the parental S. pombe strain NCYC2036 (Losson and Lacroute, 1983) to yield strains DB7 and DB4. Correct integration into the leu1 locus was verified by testing growth of clones on EMM dishes without leucine. Subsequently, strains containing an integrated UGT expression cassette were each transformed with pREP1-UGDH, pREP1-SPCC794.10, and pREP1-SPCC1322.04, respectively, to yield strains DB14, DB64, and DB65 from the parental strain DB4, and DB17, DB66, and DB67 from the parental strain, DB7, respectively. In addition, strains containing two expression cassettes for UGT2B7*1 or UGT2B7*2 were generated by transforming DB4 with pREP1-UGT2B7*2 and DB7 with pREP1-UGT2B7*1 to yield strains DB62 and DB63, respectively.
Biomass Production.
All liquid cultures lacked thiamine to induce expression by the nmt1 promoter. Incubations were performed at 30°C and 150 rpm. Ten milliliters of EMM precultures containing the appropriate supplements were inoculated with cells grown on a dish for 2 to 3 days and incubated to stationary phase; these cells were then used to inoculate 100-ml main cultures. Main cultures were incubated for 1 day for the strain NCYC2036; for 2 days for the strains DB4, DB7, DB62, DB63, DB64, DB65, DB66, and DB67; and for 5 days for the strains DB14 and DB17.
Whole-Cell Biotransformation Assay.
Whole-cell biotransformation assays were done as described previously (Dragan et al., 2010) using S-Ibu, R-Ibu, or rac-Ibu as substrate (from a 10 mM stock solution in 50% ethanol) with a final concentration of 500 μM. Afterward, culture samples were frozen until sample preparation and liquid chromatography-mass spectrometry (LC-MS) analysis. All experiments were done in triplicate.
Synthesis of Isotope-Labeled Glucuronide Metabolites.
For the synthesis of 13C6-labeled glucuronides, whole-cell biotransformation assays were done in 200-μl scale. A main culture (1.6 ml) was centrifuged (5 min, 3000g, and room temperature), washed with 1 ml of water to remove a maximum of remaining unlabeled glucose, and resuspended in EMM containing 20 g · l−1 [13C6]glucose at 200 μl. Ten microliters of a 10 mM substrate stock solution (in 50% ethanol) were added to reach a substrate concentration of 500 μM. Samples were horizontally shaken in a 2-ml reaction tube with a pinhole for 72 h at 30°C and 150 rpm. The biomass dry weight was determined from the initial main culture. All experiments were done in triplicate.
Sample Preparation.
Cell suspensions from whole-cell biotransformations were thawed if frozen and diluted 1:1 with acetonitrile. All samples (from in vitro incubations and whole-cell biotransformations) were centrifuged (5 min, 15,000g, and room temperature), the supernatant was centrifuged again (10 min, 15,000g, and room temperature), and the resulting supernatants were analyzed by LC-MS.
Liquid Chromatography-Mass Spectrometry.
Samples were analyzed on an Agilent/HP 1100 series high-performance liquid chromatograph equipped with a G1315B diode array detector (DAD) and a G1946A single quadrupole mass spectrometer (Agilent Technologies, Santa Clara, CA). Components were separated on a Lichrospher RP 18 column (particle size 5 μm, column inner diameter 4 mm, and length 125 mm; WICOM, Heppenheim, Germany) using water (mobile phase A) and methanol containing 0.1% formic acid (mobile phase B). The flow rate was 0.8 ml/min, and the column temperature was 40°C. LC-MS settings were as follows: flow profile 0 to 5 min, 10% B; 5 to 20 min, 10 to 90% B; 20 to 25 min, 90% B; 25 to 30 min, 10% B; run time 30 min; injection volume 10 μl; detector 1 G1315B DAD; DAD range 200 to 600 nm, slit width 2 nm; detector 2 G1946A mass selective detector, ionization electrospray ionization (−), and mode scan; and N2 flow rate 12 l · min−1, nebulizer pressure 45 psig (0.07 bar), nebulizer temperature 350°C, capillary voltage 3500 V, and fragmentor voltage 50 V. The quantification of metabolites was done using an external S-Ibu straight calibration line prepared in a 1:1 mixture of mobile-phase components (concentrations from 10 to 250 μM S-Ibu). Labeled metabolites were quantified using an external S-Ibu standard of 100 μM prepared in a 1:1 mixture of mobile-phase components.
Results
Human Liver Microsomes Produce Ibuprofen Acyl Glucosides but Not Galactosides.
HLMs were incubated with all possible combinations of one of the substrates (S-Ibu, R-Ibu, or rac-Ibu) and one of the cofactors (UDP-GA, UDP-Glc, or UDP-Gal). As expected, subsequent LC-MS analysis revealed that the presence of UDP-GA led to the generation of Ibu-AGs, which were identified by a m/z value of 381 for [M-H]− in negative ion mode (Fig. 1). With all three substrates, samples with UDP-Glc displayed an UV peak containing the m/z values 403 and 413, which correspond to chlorine [M+Cl]− or formic acid [M+FA-H]− adducts of Ibu acyl glycosides. Abundance values of the extracted ion currents indicate that in samples with both cofactors (i.e., UDP-GA and UDP-Glc), metabolism of S-Ibu was faster than that of R-Ibu. It was interesting to note that Ibu-AG was also present in a very low amount in samples with UDP-Glc (data not shown), which indicates the presence of a (albeit low) UGDH activity in the microsome preparation. In samples with UDP-Gal and in control reactions (without a cofactor), essentially no Ibu metabolites were detected. A scheme of the glucuronidation and glucosidation reactions of both Ibu isomers is shown in Fig. 2.
Production of Ibu-AG and Ibu-Glc by Recombinant Fission Yeast Strains That Overexpress Different Combinations of UGT2B7 and UGDH.
We previously showed that the fission yeast S. pombe is a suitable organism for the functional expression of human UGTs (Dragan et al., 2010). Because human UGT2B7 exhibits high activity toward Ibu (Sakaguchi et al., 2004; Kuehl et al., 2005) and, moreover, is able to use UDP-Glc as a cofactor for the production of a glucoside (Tang et al., 2003), we cloned UGT2B7-expressing fission yeast strains in the study presented here to examine their ability for Ibu glucosidation. Because the two polymorphic variants UGT2B7*1 and UGT2B7*2 have been shown to possess varying activities against some substrates (Thibaudeau et al., 2006; Belanger et al., 2009), we included them both in this work. Fission yeast strain NCYC2036 was transformed with the expression plasmids pCAD1-UGT2B7*1 and pCAD1-UGT2B7*2 to yield the new strains DB7 and DB4, respectively (all strains are listed in Table 1). Because (in contrast to UDP-Glc and UDP-Gal) there is no endogenous UDP-GA production in S. pombe (Dragan et al., 2010), these two strains lack the potential of glucuronide formation. For control purposes, both strains were then transformed with the plasmid pREP1-UGDH (Dragan et al., 2010) to yield DB17 (from DB7) and DB14 (from DB4). Because UGDH catalyzes the production of UDP-GA from UDP-Glc, these latter strains are capable of self-sufficient glucuronide production. In addition, UGT double-expressing strains were generated by transformation of strain DB7 with the plasmid pREP1-UGT2B7*1 and of DB4 with the plasmid pREP1-UGT2B7*2, respectively, to yield strains DB63 and DB62. All newly created strains were used for whole-cell biotransformation experiments with the substrates S-Ibu, R-Ibu, or rac-Ibu (Table 2). As expected, the parental strain NCYC2036 did not produce any Ibu metabolites and the single expressor strains DB4 and DB7 (containing UGT2B7*2 and UGT2B7*1, respectively) produced Ibu-Glc but not Ibu-AG because of a lack of UDP-GA. It was interesting to note that the Ibu-Glc production rates for the different substrate enantiomers differed significantly, with S-Ibu again being much more readily converted than R-Ibu. Furthermore, in almost all cases the two UGT2B7*1 expressing strains showed an almost doubled biotransformation rate as compared with the UGT2B7*2 expressors. In addition, the UGT double-expressing strains DB62 and DB63 did not display enhanced reaction rates in comparison to the parental single expressing strains DB4 and DB7 (data not shown), indicating that the intracellular UGT levels were not rate-limiting under these conditions. In line with expectations, the UGDH coexpressing strains DB14 and DB17 were able to produce Ibu-AG and Ibu-Glc, and biotransformation of S-Ibu was again preferred over R-Ibu. However, there was no consistent picture with respect to the relative formation of Ibu-AG versus Ibu-Glc, with glucuronide production being preferred over glucoside production in some instances but not in others.
Identification of the Fission Yeast UGPase fyu1.
We speculated that an increase in intracellular UDP-Glc levels would enhance UGT-dependent glucoside production by whole-cell biotransformation in fission yeast. UDP-Glc production depends on UGPases, a family of enzymes that catalyze the balanced reaction of the formation of UDP-Glc and pyrophosphate from UTP and glucose-1-phosphate and that were identified in many organisms (Thoden and Holden, 2007). In the genome of fission yeast, there are two putative UGPase homologs with the systematic designations SPCC794.10 and SPCC1322.04, respectively (Wood et al., 2002). We amplified the coding sequences of both genes from strain NCYC2036 and cloned each of them into the expression plasmid pREP1. Strain DB4 was then transformed with pREP1-SPCC794.10 to obtain strain DB64 and with pREP1-SPCC1322.04 for strain DB65. In the same way, DB7 was transformed with pREP1-SPCC794.10 for strain DB66 and with pREP1-SPCC1322.04 for strain DB67. Thus, these strains each coexpress a human UGT2B7 isoform and one of the two fission yeast UGPase homologs. Whole-cell biotransformations of these strains with Ibu (Fig. 3) showed that overexpression of SPCC794.10 did not lead to an enhanced Ibu-Glc production rate for any of the substrates in strains DB64 and DB66 as compared with their parental strains DB4 and DB7, respectively. Thus, these data do not support the notion that SPCC794.10 is an UGPase. In contrast, overexpression of SPCC1322.04 led to a significant increase in the production of Ibu-Glc from S-Ibu by strain DB67 in comparison to the parental strain DB7, whereas for strain DB65 no enhanced Ibu-Glc production could be detected (probably because of the weaker activity of UGT2B7*2 for this reaction). These data suggest that only one of the two putative fission yeast UGPase homologs—SPCC1322.04—indeed displays this activity. We therefore suggest to name it fyu1 for fission yeast UGPase 1.
Production of Stable Isotope-Labeled Ibu-AG and Ibu-Glc.
The availability of stable isotope-labeled reference standards is expedient for the sensitive quantification of metabolites in biological matrices by LC-MS and thus facilitates the toxicokinetic monitoring of metabolites during nonclinical safety studies. We have previously demonstrated that cultivation of UGT-expressing fission yeast strains in media containing [13C6]glucose is a convenient method of producing 13C6-labeled glucuronides (Dragan et al., 2010). In this study, whole-cell biotransformations were done using the substrate S-Ibu and strains DB14 and DB17 for the production of S-Ibu-[13C6]AG and strain DB67 for the production of S-Ibu-[13C6]Glc, respectively (Fig. 4). The comparison of the LC-MS analysis of the extracted ion currents of nonlabeled (Fig. 4, a and d) and 6-fold labeled metabolites (Fig. 4, b and e), as well as the respective fragmentation spectra (Fig. 4, c and f), unambiguously shows the successful formation of S-Ibu-[13C6]AG and S-Ibu-[13C6]Glc, respectively.
Discussion
Although glucuronide formation is a major reaction in mammalian phase II metabolism, a significant number of glucoside metabolites has also been identified, including the O-acyl glucoside of an endothelin ETA antagonist (Tang et al., 2003) and the N-glucoside of bromfenac (Kirkman et al., 1998). In addition, the formation of bile acid O-glucosides by human liver microsomes and their occurrence in human urine has been known for a long time, whereas acyl galactosides of cholic acid and deoxycholic acid were described more recently (Goto et al., 2005). Thus, it was the aim of this study to investigate the possible formation of acyl glucoside metabolites of Ibu. Biotransformation experiments with HLMs demonstrated the production of Ibu-Glc in the presence of the cofactor UDP-Glc (Fig. 1). As expected, Ibu-AG was produced in the presence of UDP-GA, whereas traces of it could also be detected in all incubations with UDP-Glc, in two incubations with UDP-Gal and one incubation without cofactor (out of nine incubations each), respectively. The detection of Ibu-AG in incubations with UDP-Glc is most likely caused by remaining UGDH activity in the microsome preparation, whereas in samples with UDP-Gal detection of Ibu-AG correlated with that of Ibu-Glc, indicating that in those two samples two further reactions took place. First, UDP-Gal was converted by UDP-Gal-4-epimerase to UDP-Glc and subsequently yielded Ibu-Glc; second, part of the UDP-Glc was converted by UGDH to UDP-GA and allowed the production of Ibu-AG. Unfortunately, quantitative analysis of the Ibu-Glc formation was not possible because of further compounds in the reaction mixture that contributed to the UV signals of the targets to an unknown extent (data not shown). Still, these data unambiguously demonstrate the conversion of R-Ibu and S-Ibu to their respective acyl glucosides by human enzymes.
Because it was known that human UGT2B7 can metabolize Ibu (Sakaguchi et al., 2004; Kuehl et al., 2005) and produce glucosides from other xenobiotics (Tang et al., 2003), it was reasonable to assume a participation of this enzyme in Ibu-Glc production. To verify this assumption, we adapted our previously established UGT expression system (Dragan et al., 2010) to the functional expression of the two polymorphic variants UGT2B7*1 and UGT2B7*2. Recombinant fission yeast strains that express one of these variants were found to be capable of producing Ibu-Glcs, whereas upon coexpression of human UGDH production of Ibu-Glcs and Ibu-AGs was observed (Table 2). The data clearly show a preference of both UGT2B7 isoforms for the isomer S-Ibu as a substrate for glucosidation and glucuronidation reactions, which is in agreement with previous studies in which a preferred glucuronidation of S-Ibu was shown in vivo (Lee et al., 1985; Tan et al., 2002) and for immobilized HLM proteins in vitro (el Mouelhi et al., 1987). However, to complicate matters, it is known that a conversion of R-Ibu to S-Ibu is accomplished by 2-arylpropionyl-CoA epimerase (Shieh and Chen, 1993; Reichel et al., 1997). A bioinformatic search did not reveal homologs of this enzyme in the genome of S. pombe (data not shown), but nevertheless it cannot be excluded that such a reaction might also occur in fission yeast. In fact, such an effect might contribute to the rather large S.D. that we observed in some of the biotransformations. In addition, an endogenous UDP-Gal-4-epimerase homolog has been described (Suzuki et al., 2010), and thus, a production of ibuprofen acyl galactoside can also not completely be ruled out, but it appears to be unlikely in view of the HLM results described above. It is interesting to note that a comparison of the activities of strains DB14 and DB17 with their parental strains DB4 and DB7 shows that in most cases the presence of UDP-GA does not significantly inhibit the Ibu-Glc production rate, indicating no strong preference of the enzymes for this cofactor. The one exception to this observation is the production of S-Ibu-Glc by UGT2B7*1, which upon UGDH coexpression is reduced to less than half of its former value. In this connection it is interesting to note that another study even observed a preference for UDP-Glc as a cofactor for a UGT2B7-dependent glucosidation (Tang et al., 2003). It will be interesting to see whether more examples of human UGTs preferring to convert some substrates with other UDP-sugars than UDP-GA will be presented in the future.
The findings of this study naturally raise the question of the physiological significance of UGT2B7-dependent acyl glucoside production. Although it is, according to current knowledge, less common than AG formation, acyl glucoside production as such appears to be relevant for endogenous substances as well as drug molecules (Stachulski, 2011). Furthermore, other groups have demonstrated before that human UGT2B7 may use UDP-Glc as a cofactor for the in vitro production of various glucosides (Mackenzie et al., 2003; Tang et al., 2003; Toide et al., 2004). Microarray data show that there are many human tissues that express UGDH to various extents (with strong signals in liver and colon), whereas UGT2B7 is predominantly expressed in liver and kidney (Yanai et al., 2005; Dezso et al., 2008). Thus, it is conceivable to envision a metabolic situation in a certain tissue (e.g., kidney) where there is strong UGT2B7 activity together with a significant level of UDP-Glc but low availability of UDP-GA. Such a situation would be expected to favor UGT2B7-dependent glucoside production. Finally, the example of the N-glucosidation of the aldose reductase inhibitor (R)-(−)-2-(4-bromo-2-fluorobenzyl)-1,2,3,4-tetrahydropyrrolo[1,2-α]pyrazine-4-spiro-3′-pyrrolidine-1,2′,3,5′-tetrone (AS-3201) demonstrated that UGT-dependent glucosidation might even take place in the presence of UDP-GA because some UGT isoenzymes (including UGT2B7) only catalyzed the glucosidation but not the glucuronidation reaction of this substrate (Toide et al., 2004). Taken together, all of this evidence may be interpreted in a way that allows for UGT2B7-dependent acyl glucoside production as a rather normal physiological process. Whether these metabolites are readily detected in blood or urine samples is another matter; in contrast to AGs, acyl glycosides do not per se contain an anion moiety and it is reasonable to assume that the dynamics of these conjugates differ from those of the respective glucuronides because in general they should differ in their ligand properties toward organic anion-transporting polypeptides or multidrug resistance-associated proteins.
With respect to Ibu metabolism, the pharmacological potency of the newly described Ibu-Glcs remains obscure, and it is conceivable that at least some of the phase I metabolites of Ibu give rise to additional acyl glucoside metabolites. A prediction of the properties of these molecules is difficult because it has been recently shown that apparently small changes in the aglycon structure may have unexpected consequences (Iddon et al., 2011). Therefore, further studies are needed to investigate the acyl migration, the hydrolysis rate, and the protein binding capacities of all of these metabolites.
From a biotechnological point of view, efficient metabolite production systems are desired to allow for the convenient production of those metabolites that are tedious to synthesize by chemical means (Zöllner et al., 2010). For the purpose of enhancing UGT2B7 biotransformation activity in our whole-cell system, we first constructed fission yeast strains that harbor two UGT expression units; however, the use of these strains did not lead to an increase in product formation as compared with the parental single-expressing strains DB4 and DB7 (data not shown). However, it stood to reason that an increase in intracellular UDP-Glc levels would facilitate UGT-dependent glucoside production. In principle, this aim might be pursued either by coexpression of a foreign (e.g., human) UGPase or by overexpression of an endogenous UGPase. In this study it was decided to attempt the latter because two putative UGPase homologs of fission yeast had already been identified by the S. pombe genome sequencing project (Wood et al., 2002). Expression plasmids containing either one of the two sequences were cloned and used to transform strains that recombinantly express UGT2B7*1 or UGT2B7*2, respectively, to yield the new strains DB64 to DB67 (Table 1). It was observed that one of two proteins—which we suggest to name fyu1—was able to significantly enhance the Ibu-Glc reaction rate, albeit only in combination with UGT2B7*1 (Fig. 3). This result suggests that, in addition to cofactor availability, UGT2B7 polymorphisms also influence the production rate of glucosides. It would be interesting to purify both proteins and determine their cofactor binding kinetics; however, these experiments were out of the scope of the study presented here. Still, it can be predicted that such an analysis will result in marked differences between the two variants.
Because the availability of stable isotope-labeled metabolites is desirable for the LC-MS analysis of biological matrices, we used our previously established general labeling technique (Dragan et al., 2010) using [13C6]glucose as a metabolic precursor that is efficiently converted within the fyu1 expressing fission yeast strains to 13C6-labeled UDP-Glc, which in turn serves as the cofactor for the UGT2B7-dependent production of S-Ibu-[13C6]Glc (Fig. 4). Likewise, when using UGDH-expressing fission yeast strains, [13C6]glucose is converted to 13C6-labeled UDP-GA and thus allows for the formation of S-Ibu-[13C6]GA. It is expected that the availability of these stable isotope-labeled metabolites will be helpful for further studies in this direction.
Authorship Contributions
Participated in research design: Bureik and Buchheit.
Conducted experiments: Buchheit and Schmitt.
Contributed new reagents or analytic tools: Dragan and Buchheit.
Performed data analysis: Buchheit and Dragan.
Wrote or contributed to the writing of the manuscript: Buchheit and Bureik.
Acknowledgments
We thank Eva Johannes for technical assistance.
Footnotes
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
doi:10.1124/dmd.111.041640.
-
ABBREVIATIONS:
- UGT
- UDP-glycosyltransferase
- AG
- acyl glucuronide
- fyu1
- fission yeast UDP-glucose pyrophosphorylase 1
- HLMs
- human liver microsomes
- Ibu
- ibuprofen
- R-Ibu
- (R)-(−)-ibuprofen
- S-Ibu
- (S)-(+)-ibuprofen
- rac-Ibu
- racemic ibuprofen
- Ibu-AG
- ibuprofen acyl glucuronide
- Ibu-Glc
- ibuprofen acyl glucoside
- UDP-GA
- UDP-α-d-glucuronic acid
- UDP-Gal
- UDP-α-d-galactose
- UDP-Glc
- UDP-α-d-glucose
- UGDH
- UDP-glucose-6-dehydrogenase
- UGPase
- UDP-glucose pyrophosphorylase
- LC-MS
- liquid chromatography-mass spectrometry
- DAD
- diode array detector
- EMM
- Edinburgh minimal medium.
- Received July 6, 2011.
- Accepted August 23, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics